modulation of intestinal permeability: an innovative method of oral drug delivery for the treatment...

MOLECULAR GENETICS AND METABOLISM 64, 12–18 (1998)ARTICLE NO. GM982667

Modulation of Intestinal Permeability: An Innovative Method ofOral Drug Delivery for the Treatment of Inherited and

Acquired Human Diseases

Alessio Fasano

Division of Pediatric Gastroenterology and Nutrition and Gastrointestinal Pathophysiology Section, Center for Vaccine Development,School of Medicine, University of Maryland, Baltimore, Maryland 21201

Received September 2, 1997, and in revised form December 1, 1997

utility of the paracellular route for oral drug deliveryConventional forms of administrations of nonab- has remained unexplored due to our limited under-sorbable drugs and peptides rely on their paren-

standing of tj physiology and the lack of substancesteral injection. The intestinal epithelium representscapable of increasing the tj permeability without irre-the major barrier to the oral absorption of theseversibly compromising intestinal integrity and functiontherapeutical agents into the systemic circulation.(1–4). Indeed, attempts to find ways to increase para-Recently, a number of innovative drug delivery ap-

proaches have been developed, including the drug cellular transport by loosening intestinal tj have beenentrapment within small vesicles or their passage hampered by unacceptable side effects induced by thethrough the intestinal paracellular pathway. Zo- potential absorption enhancing agents tested so far (1–nula occludens toxin, a recently discovered protein 4). For the most part, these agents fall within twoelaborated by Vibrio cholerae, provided tools for classes: (i) calcium chelators and (ii) surfactants (3).gaining more insights on the pathophysiology of the Both types have properties which limit their generalregulation of intestinal permeability and to devel- utility as a means of promoting absorption of variousoping alternative approaches for the oral delivery of

molecules. In the case of calcium chelators, Ca2/ deple-drugs and macromolecules normally not absorbedtion induces global changes in the cells, including dis-through the intestine. q 1998 Academic Pressruption of actin filaments, disruption of adherent junc-Key Words: Zonula occludens toxin; tight junc-tions, and diminished cell adhesion (4). In the case oftions; intestine; insulin; diabetes; IgG; oral delivery.surfactants, the potential lytic nature of these agentsmay cause exfoliation of the intestinal epithelium, irre-

In the past few years we have witnessed an explosion versibly compromising its barrier functions (3).in research aimed at creating new oral drug delivery Considering these limitations, it was reasonablesystems. This research has been fueled by unprecedent to explore whether findings from basic research onchallenges, such as the need to deliver newer and com- tj regulation can be applied to developing new ap-plex drugs (i.e., proteins, hormones, etc.) that are be- proaches to enhancing drug absorption through thecoming available through genetic engineering. Thus, paracellular route. Before addressing these issues,considerable attention has been directed at finding it is worth reviewing some of the structural and bio-ways of increasing the intestinal permeability to these chemical features of tj.compounds, with a major emphasis placed on the useof the paracellular pathway. There is now a large body

MOLECULAR COMPOSITION OF THEof evidence suggesting that tight junctions (tj) play aINTESTINAL TIGHT JUNCTIONSpivotal role in epithelial permeability. However, the

The intestinal epithelium represents the largestSupported by the National Institute of Health Grants DK-48373 and AI-35740. interface (more then 2,000,000 cm2) between the ex-

121096-7192/98 $25.00Copyright q 1998 by Academic PressAll rights of reproduction in any form reserved.

AID MGM 2667 / af13$$$221 06-26-98 10:12:37 bmmal AP: MGM

13ORAL DRUG DELIVERY

erodimer (17) in a detergent-stable complex with anuncharacterized 130-kDa protein (ZO-3). Most im-munoelectron microscopic studies have localized ZO-1 to precisely beneath membrane contacts (18). Twoother proteins, cingulin (19) and the 7H6 antigen(20) are localized further from the membrane andhave not yet been cloned. Rab 13, a small GTP bind-ing protein has also recently been localized to thejunction region (21). Other small GTP-binding pro-teins are known to regulate the cortical cytoskeleton;i.e., rho regulates actin-membrane attachment in fo-cal contacts (22), and rac regulates growth factor-induced membrane ruffling (23). Based on the anal-ogy with the known functions of plaque proteins inthe better characterized cell junctions, focal contacts(24), and adherens junctions (25), it has been hy-pothesized that tj-associated plaque proteins are in-volved in transducing signals in both directions

FIG. 1. Model for components of the tight junction. Occludin, across the cell membrane and in regulating links tothe transmembrane protein strand, is anatomically and function-the cortical actin cytoskeleton.ally connected with the cell cytoskeleton via the junctional com-

plex. This complex comprises a series of proteins, including ZO-1, ZO-2, and p130 (ZO-3). Other proteins, such as cingulin, 7H6, REGULATION OF INTESTINAL TIGHTrab13, rho, and ras, are located further from the cell membrane. JUNCTIONSHowever, they seem also involved in the regulation of tight junc-tion permeability.

To meet the many diverse physiological and patho-logical challenges to which epithelia are subjected, tjmust be capable of rapid and coordinated responsesternal environment and the internal host milieu and

constitutes the major barrier through which mole- that require the presence of a complex regulatory sys-tem. The precise characterization of the mechanismscules can be either absorbed or secreted. There is

now substantial evidence that tj play a major role involved in the assembly and regulation of tj is an areaof current active investigation. The discovery of Zonulain regulating epithelial permeability by influencing

paracellular flow of fluid and solute. Moreover, struc- occludens toxin (Zot) elaborated by Vibrio cholerae(26,27) shed some light on the intricate mechanismstural features of occluding junctions such as strands

number often correlate inversely with the perme- involved in the regulation of tj permeability. As it oftenoccurs in science, the discovery of this protein wasability of epithelia as measured electrophysiologi-

cally (5). A century ago, tj were thought to be a se- made by accident. A few years ago, researchers at theCenter for Vaccine Development at the University ofcreted extracellular cement forming an absolute and

unregulated barrier within the paracellular space Maryland in Baltimore engineered what was believedto be an ideal attenuated vaccine for cholera. At that(6). There is now abundant evidence that tj are dy-

namic structures that readily adapt to a variety of time cholera toxin (CT) was the only described‘‘weapon’’ used by V. cholerae to induce diarrhea.developmental (7,8), physiological (9–11), and patho-

logical (12–14) circumstances. Therefore, the deletion of the gene encoding the activesubunit of CT appeared to be the best approach toThe assembly of tj is the result of cellular interac-

tions that trigger a complex cascade of biochemical eliminate the key pathogenic factor of the microorgan-ism while maintaining the expression of other Vibriosevents that ultimately lead to the formation and

modulation of an organized network of tj elements, antigens necessary for a protective immune response.When fed to volunteers, these vaccine candidates stillthe composition of which has been only partially

characterized (15) (Fig. 1). A candidate for the trans- caused mild diarrhea in more than one-half of the vac-cinees (28). In search of other factors responsible formembrane protein strands, occludin, has recently

been identified (16). Six proteins have been identi- this residual diarrhea, our group identified Zot, a pro-tein elaborated by V. cholerae that increases the per-fied in a cytoplasmic submembranous plaque under-

lying membrane contacts, but their function remains meability of the small intestine by affecting the struc-ture of tj (26).to be established (15). ZO-1 and ZO-2 exist as a het-


14 ALESSIO FASANO

FIG. 2. Proposed Zot intracellular signaling leading to the opening of intestinal tight junctions. Zot interacts with a specific surfacereceptor (1) whose distribution within the intestine varies. The protein is then internalized and activates phospholipase C (2) thathydrolyzes phosphatidyl inositol (3) to release inositol 1,4,5-Tris phosphate (PPI-3) and diacylglycerol (DAG) (4). PKCa is then activated(5), either directly (via DAG) (4) or through the release of intracellular Ca2/ (via PPI 3) (4a). PKCa catalyzes the phosphorylation oftarget protein(s), with subsequent polymerization of soluble G actin in F actin (7). This polymerization causes the rearrangement of thefilaments of actin and the subsequent displacement of proteins (including ZO1) from the junctional complex (8). As a result, intestinaltight junctions become looser.

We have subsequently demonstrated that Zot acti- addition of the protein to the intestinal mucosa andis readily reversible once the toxin is removed.vates a complex intracellular cascade of events that

regulate the intestinal permeability (Fig. 2) (29). Zot The discovery of Zot raised an obvious question:induces a dose- and time-dependent PKCa-related Why does a microorganism that elaborates the mostpolymerization of actin filaments strategically local- powerful enterotoxin ever described (NB: 15 mg ofized to regulate the paracellular pathway (29). These purified cholera toxin fed to human volunteers in-changes are a prerequisite to the opening of tj and duced as much as 24 liters of diarrhea) maintain inare evident at a toxin concentration as low as 1.1 1 its limited genome genes encoding other enterotox-10013 M (30). The toxin exerts its effect by interacting ins, such as Zot? We do not have a definitive answerwith a specific surface receptor that is present on yet; however, it is possible to speculate that the elab-mature cells of small intestinal villi, but not in the oration of different toxins may ensure the conserva-

tion of V. cholerae species. Diarrhea has always beencolon (30). The regional distribution of Zot recep-tor(s) coincides with the different permeabilizing ef- considered a defense mechanism of the host animal

against V. cholerae infection. However, a differentfect of the toxin on the various tracts of intestinetested (30). Both in vivo (30,31) and in vitro way of interpreting this phenomenon should be con-

sidered. The ideal habitat for V. cholerae is repre-(26,30,31) studies demonstrated that the effect of Zoton tissue permeability occurs within 20 min of the sented by the aquatic environment where they usu-



FIG. 3. Reversible effect of purified Zot on tissue resistance (A and B) and transepithelial transport of insulin (C) and IgG (D) in rabbitileum in vitro. Paired tissues, matched on the basis of their resistance, were exposed luminally to either [125I]insulin 10011 M (2 mCi Å 10012

M) (left) or [125I]IgG 156.25 ng (1 mCi Å 83.3 ng) (right), alone (s) or in the presence of 1.1 1 10010 M Zot (h). After 80 min of incubation,the Ringer’s solutions were replaced with solutions of identical composition but without Zot. Zot reversibly increased the transepithelialabsorption of both insulin and IgG. These effects paralleled the Rt decrement induced by the toxin. Number of animals Å 4. Reproduced fromThe Journal of Clinical Investigation, 1997, 99, pp. 1158–1164, by copyright permission of The American Society for Clinical Investigation.

ally live and where temperature, pH, and O2 concen- mucosa. Zot: (a) is not cytotoxic and does not affectthe viability of the intestinal epithelium ex vivotration are ideal for their growth and replication.(26,29); (b) fails to completely abolish the intestinalThe ingestion by a mammalian host represents atransepithelial resistance (26,29,31); (c) interactsbiological drama for V. cholerae. The microorganismwith a specific intestinal receptor whose regional dis-has to face a very hostile environment and the onlytribution within the intestine varies (30); (d) is notway to survive is to escape from the intestine andeffective in the large intestine where the presence ofto return to its aquatic reservoir. Since Vibrios arethe colonic micro flora could be potentially harmful ifexcellent swimmers, water represents the only vehi-the mucosal barrier was compromised (30,31), i.e.,cle by which to return to the natural habitat. There-does not induce acute systemic side-effects (for atfore, the elaboration of several factors that, by differ-least 80–90 h) when orally administered (31) and,ent mechanisms of action, all may induce fluid accu-lastly; (f) induces a reversible increase of tissue per-mulation within the intestinal lumen, represents ameability (26,29,31).great advantage for V. cholerae.

To establish the efficacy of Zot as an intestinalabsorption enhancer, we selected insulin and immu-USE OF ZOT AS A TOOL FOR ORALnoglobulin G (IgG). This choice was based on theDRUG DELIVERYrelative size and structure, biological activities, and

Zot displays multiple properties that make it the therapeutic relevance of these proteins. In vitro ex-most promising tool currently available to enhance periments in the rabbit ileum mounted in Ussing

chambers demonstrated that Zot (1.1 1 10010 M) re-drug and peptide transport through the intestinal


16 ALESSIO FASANO

versibly increases the intestinal absorption of bothinsulin (by 72%) and IgG (by 52%) in a time-depen-dent manner (31). Zot permeabilizing effect peakedat 80 min and was completely reversible within 20min of the withdrawal of the toxin from the Ussingchambers (Fig. 3). This Zot-induced increase in ab-sorption coincided with a reduction in tissue resis-tance (Rt) (Fig. 3). When tested in the intact host byusing the rabbit in vivo perfusion assay, Zot (1.1 110010 M) increased the passage of insulin across boththe jejunum and distal ileum 10-fold, whereas nosubstantial changes were observed in the colon (31)(Fig. 4). The increased absorption of insulin was re-ciprocal with a shift of water absorption toward se-cretion (Fig. 4), a change that has been related tothe permeabilizing effect of Zot on the paracellularpathway in vivo (30). This effect was detectable assoon as 20 min after Zot perfusion in the small intes-tine and was completely reversible within 60 min ofits withdrawal (Fig. 4). Zot also reversibly increasedthe serum concentration of both insulin and the non-absorbable marker [14C]polyethylene glycol (PEG)-4000 from jejunum and ileum, but not from colon(31). Similar results were obtained with IgG,whereby Zot (1.1 1 10010 M) induced twofold andsixfold increases of IgG absorption in the jejunumand ileum, respectively. Again, no increases in ab-sorption were detected in the colon (31).

To evaluate the bioactivity of insulin after enteralcoadministration with Zot, the hormone was orallyadministered to acute type 1 diabetic male BB/Worrats with or without Zot, and the blood glucose levelsof the rats were serially measured. After oral admin-istration of insulin alone, given at doses between 5and 30 IU, blood glucose levels of treated animalswere not appreciably lowered (31). In contrast, wheninsulin at doses as low as 10 IU was orally coadmin-istered with Zot 1.1 1 10010 mol (5 mg), a significantreduction in blood glucose concentration was ob-served (Fig. 5). This decrement was comparable to FIG. 4. Effect of purified Zot on water (j) and insulin (h)that seen with a conventional dose of SQ insulin and transport, as determined by the in vivo perfusion assay, in rabbit

jejunum, distal ileum, and colon. Note the reversible incrementreturned to baseline by 6 h postadministration (Fig.of insulin absorption that Zot induced in the small, but not in the5). None of the animals treated with insulin / Zotlarge intestine. This effect coincided with the decreased absorp-experienced fever or other systemic symptoms, andtion of water evoked by the toxin. Reproduced from The Journal

no structural changes could be demonstrated in the of Clinical Investigation, 1997, 99, pp. 1158–1164, by copyrightsmall intestine on histological examination (31). permission of The American Society for Clinical Investigation.Furthermore, Zot administration did not induce di-arrhea, despite the secretory effect of the toxin. Thelack of diarrhea is probably related to the distribu- gradient. The excess of fluid that accumulates in the

small intestine is completely reabsorbed in the colontion of the Zot receptors within the intestine (30).Following the activation of the Zot receptors in the (where tj regulation is not operative because of the

lack of Zot receptors), preventing intestinal fluid losssmall intestine, tj are reversibly opened and fluidleaks into the intestinal lumen driven by the osmotic and, therefore, diarrhea.



normally not absorbed through the intestine. Thepromising results obtained in the animal model,both in vitro and in vivo, represent an encouragingbasis for further studies to establish the possibleclinical applications of this system for the treatmentof human diseases that currently require frequentand long-life parenteral drug administration.

ACKNOWLEDGMENTS

Figures 3, 4, and 5 were reproduced from the Journal of ClinicalInvestigation, 99: 1158–1164, 1997, by copyright permission ofthe American Society for Clinical Investigation.

REFERENCES

1. Lee VHL, Yamamoto A, Kompella VB. Mucosal penetrationenhancers for facilitation of peptide and protein drug absorp-tion. Crit Rev Ther Drug Carrier Sys 8(2):91–192, 1991.

2. Muranishi S. Absorption enhancers. Crit Rev Ther DrugCarrier Sys 7(1):1–33, 1990.

FIG. 5. Effect of oral insulin 10 IU, alone (j) or in the pres- 3. Hochman J, and Artursson P. Mechanisms of absorptionence of Zot 5 mg (h) on serum glucose decrement in BB/Wor dia- enhancement and tight junction regulation. J Controlled Re-betic rats. The coadministration of Zot induced a reduction in lease 29:253–267, 1994.blood glucose concentration comparable to that seen with a con- 4. Citi S. Protein kinase inhibitors prevent junction dissocia-ventional dose of SQ insulin (l) and returned to baseline by 6 h tion induced by low extracellular calcium in MDCK epithe-postadministration. Blood glucose decrement of untreated ani- lial cells. J Cell Biol 117(1):169–178, 1992.mals (s) and animal treated with oral Zot alone (m) are shown

5. Madara JL, Dharmsathaphorn. Occluding junction struc-for comparison. Number of observations Å 3. *P Å 0.005; ‡P Åture-function relationship in a cultured epithelial mono-0.003; §P Å 0.009, as compared to that seen with oral insulinlayer. J Cell Biol 101:2124–2133, 1985.alone. Reproduced from The Journal of Clinical Investigation,

6. Cereijido M. Evolution of ideas on the tight junction. In Tight1997, 99, pp. 1158–1164, by copyright permission of The Ameri-Junctions. Boca Raton, FL:CRC pp 1–13, 1992.can Society for Clinical Investigation.

7. Magnuson T, Jacobson JB, Stackpole CW. Relationship be-tween intercellular permeability and junction organizationin the preimplantation mouse embryo. Dev Biol 67:214–224,Taken together, these results demonstrate that 1978.

coadministration of Zot with biologically active in-8. Schneeberger EE, Walters DV, Olver RE. Development of

gredients enhances intestinal absorption of the ac- intercellular junctions in the pulmonary epithelium of thetive molecule, and that this enhancement is effective foetal lamb. J Cell Sci 32:307–324, 1978.for both relatively small (5733 Da: insulin) and large 9. Madara JL, Pappenheimer JR. Structural basis for physio-

logical regulations of paracellular pathways in intestinal ep-molecules (140–160 kDa: IgG). Furthermore, the ex-ithelia. J Membr Biol 100:149–164, 1987.periments in BB/Wor diabetic rats demonstrate that

10. Mazariegos MR, Tice LW, Hand AR. Alteration of tight junc-orally delivered insulin can retain its biological ac-tional permeability in the rat parotid gland after isoprotera-tivity without provoking severe hypoglycemia withinnol stimulation. J Cell Biol 98:1865–1877, 1984.

the range of the insulin administered, i.e., up to 1511. Sardet C, Pisam M, J. The surface epithelium of teleostean

times more than the effective parenteral insulin fish gills. Cellular and junctional adaptations of the chloridedose. These findings have important practical impli- cell in relation to salt adaptation. J Cell Biol 80:86–117,

1979.cations, since the insulin therapeutic index (i.e., theratio between the median toxic dose and the median 12. Milks LC, Conyers GP, Cramer EB. The effect of neutrophil

migration on epithelial permeability. J Cell Biol 103:2729–therapeutic dose) is relatively low.2738, 1986.

13. Nash S, Stafford J, Madara JL. The selective and superox-CONCLUSIONSide-independent disruption of intestinal epithelial tightjunctions during leukocyte transmigration. Lab Invest

Current knowledge on the regulation of intestinal 59:531–537, 1988.tight junctions by Zonula occludens toxin were ap- 14. Shasby DM, Winter M, Shasby SS. Oxidants and conduc-

tance of cultured epithelial cell monolayers: Inositol phos-plied to enhance the absorption of macromolecules


18 ALESSIO FASANO

pholipid hydrolysis. Am J Physiol 255 (Cell Physiol 24): 24. Guan JL, Shalloway D. Regulation of focal adhesion-acti-vated protein tyrosin kinase by both cellular adhesion andC781–C788, 1988.oncogenic transformation. Nature 358:690–692, 1992.15. Anderson JM, Balda MS, Fanning AS. The structure and

regulation of tight junctions. Cell Biol 5:772–778, 1993. 25. Tsukita S, Itoh M, Nagafuchi A, Yonemura S, Tsukita S.16. Furuse M, Hirase T, Itoh M, Nagafuchhi A, Yonemura S, Submembraneous junctional plaque proteins include poten-

Tsukita S, Tsukita S. Ocludin: A novel integral membrane tial tumor suppressor molecules. J Cell Biol 123:1049–1053,protein localizing at tight junctions. J Cell Biol 123:1777– 1993.1788, 1993. 26. Fasano A, Baudry B, Pumplin DW, Wasserman SS, Tall BD,

17. Gumbiner B, Lowenkopf T, Apatira L. Identification of 160- Ketley JM, Kaper JB. Vibrio cholerae produces a secondkDa polypeptide that binds to the tight junction protein ZO- enterotoxin, which affects intestinal tight junctions. Proc1. Proc Natl Acad Sci USA 88:3460–3464, 1991. Natl Acad Sci USA 88(22):5242–5246, 1991.

18. Stevenson BR, Anderson JM, Bullivant S. The epithelial 27. Baudry B, Fasano A, Ketley JM, Kaper JB. Cloning of atight junction: Structure, function, and preliminar biochemi- gene (zot) encoding a new toxin produced by Vibrio cholerae.cal characterization. Mol Cell Biochem 83:129–145, 1988. Infect Immun 60(2):428–434, 1992.

19. Citi S, Sabannay H, Jakes R, Geiger B, Kendrich-Jones J.28. Levine MM, Kaper JB, Herrington D, Morris JG, ClementsCingulin, a new peripheral component of tight junctions.

ML, Black RE, Tall B, Hall R. Volunteer studies of deletionNature 333:272–275, 1988.mutants of Vibrio choleraeO1 prepared by recombinant tech-

20. Zhong Y, Saitoh T, Minase T, Sawada N, Enomoto K. Mono- niques. Infect Immun 56:161–167, 1988.clonal antibodies 7H6 reacts with a novel tight-junction as-

29. Fasano A, Fiorentini C, Donelli G, Uzzau S, Kaper JB, Mar-sociated protein distinct from ZO-1, cingulin and ZO-2. Jgaretten K, Ding X, Guandalini S, Comstock L, GoldblumCell Biol 120:477–483, 1993.SE. Zonula occludens toxin modulates tight junctions21. Zahraoui A, Joberty G, Arpin M, Fontaine JJ, Hellio R, Tavi-through protein kinase C-dependent actin reorganization,tian A, Louvard D. A small rab GTPase is distributed inin vitro. J Clin Invest 96:710–720, 1995.cytoplasmatic vesicles in non polarized cells but co-localized

30. Fasano A, Uzzau S, Fiore C, Margaretten K. The enterotoxicwith the tight junction marker ZO-1 in polarized epithelialeffect of Zonula occludens toxin on rabbit small intestinecells. J Cell Biol 124:101–115, 1994.involves the paracellular pathway. Gastroenterology22. Ridley AJ, Hall A. The small GTP-binding protein rho regu-112:839–846, 1997.lates the assembly of focal adhesions and actin stress fibers

in response to growth factors. Cell 70:389–399, 1992. 31. Fasano A, Uzzau S. Modulation of intestinal tight junctionsby Zonula occludens toxin permits enteral administration of23. Ridley AJ, Paterson HF, Johnston CL, Diekmann D, Hall A.

The small GTP-binding protein rac regulates growth factor- insulin and other macromolecules in an animal model. JClin Invest 99:1158–1164, 1997.induced membrane ruffling. Cell 70:401–410, 1992.


modulation of intestinal permeability: an innovative method of oral drug delivery for the treatment...

Documents